Looking at an Eng dwg of a GWR King I have pinned up at work I note that the pipe from the sand box under the cab goes down close to the rail-head behind the rear driver. The question I have is: how effective is this at getting sand under the wheel, as it looks like the arrangement would always be depositing sand behind the wheel, which would not then be 'run over' by the driving wheel unless the loco rolled back. Wouldn't it be better to blast sand in front of the wheel, so it gets trapped as the wheel rolls forward onto the rail? Does it rely on sand being picked up on the rear of the wheel and carried round as the wheel turns to eventually provide adhesion when that part of the wheel contacts the rail? there must be a reason?